[This is a conference report I am obliged to write for my NTNU colleagues; it is shared here in solidarity with colleagues from other institutions who are similarly obliged that they might save time/effort when reporting. As a consequence, this report may not be that interesting for the general reader]
I had the immense pleasure of attending ELAG again this year, I met with like minds and learnt a whole load of things and had time to reflect on ideas and strategies catalysed by what is happening outside the NTNU bubble.
ELAG conferences are always brilliant for people like me — lots of people who do the same kind of thing as me gathered together, many of whom I know very well and many of whom I want to get to know better. Brilliant. If you work with library technology and like talking about it, go there.
Following standard ELAG procedure, this year’s talks all centred around the idea of “lingering gold”; this might need explaining, here were talking about the value that lies in our data that we have yet to extract. As a topic, it was fair enough; did everyone stick to the topic? No. Is that OK? Yes.
The conference was in Bath, which is a nice place, albeit a bit remote given the general problem of getting to non-Schengen areas with internal travel on top. The conference was hosted by the University of Bath in new facilities; organisation was top-notch and one was met by typical-for-the-UK hospitality. Wifi was unusually good for a UK conference, while the absence of electricity outlets lead many to believe that the new building was either kitted out with some new-fangled induction charging in the desks, or was built in anticipation of some technology that would render current battery technologies obsolete.
If you went there for the food…well, the UK.
My general impression of conferences these days remains the same — ELAG was a good experience content-wise compared to the vast majority of conferences I have attended. What sticks in my mind is that the majority of the useful ideas came not from presentations, but from discussions outside the presentations both at the venue and in more informal settings over beer.
What struck me was the level of maturity thinking is reaching regarding how technology is important, but organisation is key. Linked data will only function if it has a sound fundament in the organisation. We have known this for a long time at NTNU, but it seems to be common knowledge elsewhere now. What was previously assumed to be a function of technology — say persistence — is now clearly understood to be a function of the organisation and planning. You can’t just rely on a third party — W3C or OCLC — to cover this base.
It’s also becoming clear that there’s a maturity in attitudes to development; while we might joke about the hipster coders, there is room for diverging platforms because we communicate using standardised protocols. This will hopefully help us move on from the discourse of “fuck you” common in some programming circles.
Personally, this was a conference that I will remember well because I had a very good time. I met with old friends like Lukas, Jane, Adrian, Felix, Owen, Martin, Niklas, Kristin, Peter, Joachim, Uwe and old Twitter friends like Céline, Knut, Karen, Tom, Ben, Toke and co. (not mentioned? Don’t be downhearted — I am hopeless with names) for had banter and serious discussion. The value in networking can’t be underestimated. ELAG brought people together in a very positive way that has been absent for a lot of the time in the last few years. Maybe the technology platform has moved on and optimism has returned, maybe ELAG initiated this. We’re somewhere on a curve, and that place is way better than the place we were twelve months ago.
If you’re interested in a breakdown of the presentations, I have tried to cover everything below.
Day one: Bootcamps
I attended a bootcamp/hackathon on OCLC services. This was rather more oriented towards traditional API-based services than I had hoped; I was really interested in the RDF-based services. A lot of time was spent on getting the API-specific handshaking to work with the different client libraries (and more generic tools). This kind of thing is very familiar to anyone who has worked with key based APIs previously, and is both tedious and a strong reminder why we use linked data by preference these days. That said, the content of the bootcamp was good, and the host (Steve Meyer) was skilled and gave a good overview of OCLC technologies.
If you don’t pay for OCLC things, well, it isn’t that relevant, which is a shame. It became quite apparent that I needed to concentrate on other things. Luckily, I was sat next to OCLC’s Richard Wallis, who is a linked data guy, and who gave me a few heads up regarding OCLC’s LOD services that I really appreciated. All-in-all a good session that I got a lot out of, even if I wasn’t doing what was on the schedule. Hackathons allow space for this kind of flexible learning.
The keynote by Stella Wisdom was a bit lacklustre for me; I think that it was a good presentation of the things that they’re doing at the British Library, but it wasn’t technical enough and in the context of ELAG seemed more like a sales presentation even though they aren’t selling anything. That was a shame as there are a whole host of things to learn from the British Library; no criticism can be levelled at Stella, as she replaced the invited speaker who recently changed jobs.
In the next session, Karen Coyle, Lukas Koster, Martin Malmsten, Anders Söderbäck and myself presented our allegory on library linked data. This was intended as a humourous view of some of the current issues facing libraries who do linked data.
Niklas Lindström and Lina Westerling gave a very clear presentation of how libraries should be doing things: copying development at Libris. Arrogant bastards, who take all the good ideas. I can hardly be bothered heaping more praise on them. Take a look for yourself. 😉
After a typically British lunch (no beer though), we had workshops. I attended the Jane and Adrian Stevenson’s Linking Data with sameAs: Challenges and Solutions, which introduced tools for linking data; using reconciliation services in LODrefine and SILK. Jane and Adrian gave a good introduction to these tools; while the workshop might be on the steeper end of the learning curve, I suspect that this particular curve is extremely steep for anyone not previously used to working with data in directly in different ways (XML-stack, SQL, etc.) Nevertheless, I noticed novices and seasoned pros were getting useful tips and information.
Felix Ostrowski presented development of Regal, a repository for electronic documents and bibliographic data. The presentation was well appreciated because it was untypically warts-and-all. While I am aware of several projects within this domain, Felix’s honesty regarding technology and architectural choices made this presentation doubly interesting. Speaking with Felix during breaks made it quite clear that his head is well screwed on and that the thinking in the project has helped develop better thinking for future projects.
Naeem Muhammad and Sam Alloing presented MIF: Metadata Interoperability Framework, which is in essence yet another ETL tool, albeit specifically oriented towards people working with GLAM data (it supports Europeana’s EDM). I have seen many such tools, and I still struggle so see the value in a tool that does ETL in an abstract way — almost as if understanding what ETL actually involves is pointless. Maybe I’m being harsh here; I’m not the audience for this, but there is possibly a need — filling a gap where a real-world system isn’t possible because contracts have been signed and staff have to make do. But that way sadness lies — you get to breath life into dead systems by gaffatape and bodgery to work around them.
Day three started with lightning talks, which were all very interesting; I like short, concise presentations of small, relevant things. Did I follow? I can say no because I was stressed by doing my own talk …eek.
Tamar Sadeh from Ex Libris gave a presentation title Optimisation of known item discovery, which seemed to me more like a product presentation of Ex Libris’ relevance-oriented software. This was a bit of a disappointment — I work at an Ex Libris library and their Primo software measures equally against our Ebsco Discovery Services implementation. I’m afraid I remain unconvinced that relevance is really relevant in [pun?] in relation to library metadata; I think a lot more work needs to go into the way metadata is interpreted — and it must be interpreted — before it can be relevanced.
Valentine Charles’ presentation of Discovering libraries’ gold through collections-level descriptions got me thinking more about EU-funded projects. This wasn’t Valentine’s fault, a great person, but the whole too-little-too-late nature of FP7 money has made me a sad boy and it gets a bit much.
Steve Meyer from OCLC gave a sensible presentation about building usable and useful APIs that wasn’t OCLC-oriented; all power to him for that. The points he made are common sense and should be natural for serious developers. Always nice to have this kind of thing reiterated.
Marina Muilwijk from Utrecht gave a presentation titled The black box opens which outlined how Utrecht is developing new services in new ways. This was a nice overview presentation that made me think back to UBiT2010 days. Sound information here.
A British lunch (still no beer) was followed by the second half of the workshop with Jane & Adrian. Again, sound stuff.
Peter Mayr’s presentation on EuropeanaBot reminded me that I’m also a bit of a negative Norris when it comes to bots. I guess I am a child of my time. I’m not sure I see the value from any perspective.
Stina Johansson from Chalmer’s gave a presentation that I have seen bits of before at an event in Sweden, however, things have developed and moved on, so it is all good. A nice presentation of using lingering gold from the library both in the form of data and staff skills.
I enjoyed the conference dinner at the Assembly rooms, talking at length with several people I know and some I got to know. Particularly enjoyed talking with Owen Stephens & Céline Carty who joined Lukas Koster & myself for after-event chat.
Day four: Ben O’Steen gave a fine presentation about yet more things that are happening at the British Library Labs. They obviously do a lot of quick & dirty work to test ideas, and they seem to get good results. I think we can learn from this…again.
Ciaran Talbot and Kay Munro gave a presentation on library game, which brought forward everything I hate about gamification. I really cannot stand it. I am a reactionary old fart in this respect and I feel I should apologise. I am fortunately not the target audience for gamified stuff, so I guess that is OK. The presentation, however, covered the process and thinking in the process, iterations and feedback. This was a good presentation, with some informed ideas related to something I consider to be often rather ‘pop’.
Tom Demeranville gave a clean, solid presentation on ORCID, explaining concepts and how it can be integrated into other systems and used in general. Tom’s open & dynamic style lent itself well to this rather dry topic; adding sound opinion on top of technical and implementational detail. Even though we’re talking FP7 here, I didn’t want to chew my own tongue off. Win.
Demmy Verbeke gave a rather academic presentation of work taking place at KU Leuven in support of digital humanties.
Toke Eskildsen gave a good presentation about how outsourcing (and probably in-housing) scanning isn’t plain sailing, and how they recognize tricksy hobbitses in the scanned files in Aarhus. Good technical information nicely presented.
At this point my head bluescreened and I gave up. Four days of information overload did it for me.